A framework for modeling the appearance of 3D articulated figures

Author

Sidenbladh, Hedvig ; De La Torre, Fernando ; Black, Michael J.

Author_Institution

CVAP, R. Inst. of Technol., Stockholm, Sweden

fYear

2000

fDate

2000

Firstpage

368

Lastpage

375

Abstract

This paper describes a framework for constructing a linear subspace model of image appearance for complex articulated 3D figures such as humans and other animals. A commercial motion capture system provides 3D data that is aligned with images of subjects performing various activities. Portions of a limb´s image appearance are seen from multiple views and for multiple subjects. From these partial views, weighted principal component analysis is used to construct a linear subspace representation of the “unwrapped” image appearance of each limb. The linear subspaces provide a generative model of the object appearance that is exploited in a Bayesian particle filtering tracking system. Results of tracking single limbs and walking humans are presented

Keywords

face recognition; filtering theory; image representation; principal component analysis; tracking; user interfaces; 3D articulated figures; 3D data; Bayesian particle filtering tracking; HCI; limb image; linear subspace representation; modeling; motion capture system; unwrapped image appearance; walking humans; weighted principal component analysis; Animals; Bayesian methods; Biological system modeling; Ear; Electrical capacitance tomography; Filtering; Humans; Nonlinear filters; Particle tracking; Principal component analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Automatic Face and Gesture Recognition, 2000. Proceedings. Fourth IEEE International Conference on

Conference_Location

Grenoble

Print_ISBN

0-7695-0580-5

Type

conf

DOI

10.1109/AFGR.2000.840661

Filename

840661